kn@ is working on moving python2 ports to python3 and asked me why we have an outdated version of mercurial. He wasn't the only dev asking recently.
A few years ago I decided to remove mercurial third party extensions because nobody was maintaining them, were broken for a year (or maybe even more) and nobody had complained. I kept tortoisehg because at that time upstream was releasing regularly new versions in sync with mercurial versions. Since a few mercurial releases ago, they're not releasing new versions. Tortoise uses the internal API of mercurial, so I can't update mercurial without updating also tortoisehg. We're stuck now on the version 5.0 and the mercurial devs are going to release 5.4 the next month. What are we going to lose if we keep the mercurial port stuck on 5.0?. Close to 100% python3 support, a new official extension to work with git servers and more rust code (which I'll not enable by default but I would like to test). Also, the mercurial developers are a very small team and don't release backports for security issues, so the more outdated our mercurial version is, the harder is to backport the patches for me. So, I would like to remove tortoisehg and update mercurial to the latest version. The plan B would be to create a new port with the outdated version of mercurial and update the canonical port to the latest version. The problem of that idea is that maybe eventually we will delete python2 and tortoise will be lost anyway. The only thing which users are going to lose is the graphical DAG / commits history. The package mercurial-x11 includes an extension which shows the same graphical output and you can always use "hg serve" to see the DAG / history in your browser. I would like to remove the port before of 6.7. Any objection?. Is anyone yelling at me?. I can't hear you. -- Juan Francisco Cantero Hurtado http://juanfra.info
